How to set up the ESP32-S3 AI Smart Speaker as an assistant?

[vocal sparrow]

Hi folks. Very new to Home Assistant so please bear that in mind. I run a home server (Ubuntu Desktop) with an intel Arc GPU. I'm trying to get the ESP32-S3 AI Smart Speaker up and running as a amazon replacement. Here's the site for the unit:

https://www.waveshare.com/esp32-s3-audio-board.htm

I thought AI would get me set up and running. Silly me. However, it did get me to the point where the AI model is running successfully on my gpu. I can query it successfully via the text prompt in HA. It also got me so far with setting up the ESP unit. However, I'm now totally lost and AI lost the plot leaving me hanging.

I have succesfully installed piper, faster-whisper and openwakeword as a service. Am running Home Assistant in a docker container.

In case relevant, it had me set up the esp unit with this yaml:

https://pastebin.com/Z4Lx32TS

It complied to bin fine and I could flash the unit with the bin file. HA picks it up and I can add it as a device. However, I'm not sure the wake word is set up properly as it's greyed out (see image).

I had a scout around for guides, but I swear youtube is getting deliberately s**t unless you subscribe to premium (which I refuse to).

Can anyone help or point me towards a relevant guide? This whole ecosystem is new to me (ESP, HA) so working blind.

[safe flame]

Voice stuff has advanced pretty quickly so guides and AI get outdated pretty quickly.
your ESPHome config is a bit bare.

I think that you might be better off asking this over on the ESPHome discord server.
https://discord.gg/kzQypsks

There is a good chance someone over there has an example config.

[vocal sparrow]

Thanks. Have posted over there. not sure if I'm going to get a response though!

[vocal sparrow]

Found an example config. Either it doesn't work or I've applied it wrong. Hmmm - this is quite a difficult area to make progress in isn't it?

[grand gale]

if you're still having trouble, i found a config